An Advanced Regenerative Circuit in 1953
The attached scans show an article about "An advanced regenerative circuit" by Irving Gottlieb W6HDM, from the May 1953 issue of "Radio and TV News" magazine.
The author designed an improved regenerative circuit with the goal to reduce the loading effects of grid detection. Removing detection loading by the grid circuit is particularly useful in a regenerative circuit. because loading losses are signal dependent and affect the overall regeneration.
The author replaced grid detection with anode bend detection to eliminate loading by the forward bias in grid leak detection. He also took the further step of buffering the RF signal with one extra triode to further reduce any capacitive loading by the grid circuit of the first triode.
The plate is loaded with a RF bypass capacitor, so Miller capacitance feedback through the plate-grid capacitance is eliminated in the first triode.
The necessity of the second triode as a buffer is not entirely obvious, because it might have been enough to simply tap the RF feedback coil from the cathode of the first triode. Perhaps the reduction in grid capacitance that is afforded by very lightly loaded operation of the first triode is helpful in regeneration stability.
The author uses a potentiometer at the first triode plate to control two factors of the operation. One is the location of the anode bend or cutoff point that is needed for plate detection, and the other is the follower gain of the tube that is also controlled by the plate voltage or current.
A quick look at the family of curves of the 6SN7 dual triode suggests that the cathode bias of the first triode should be a few Volts, perhaps 3V. The actual voltage depends on the adjustable plate voltage. The text suggests that optimum operation occurs with a plate voltage between 20V and 50V.
The mu, or intrinsic voltage gain of the 6SN7 is 20. This means that the cathode bias must be slightly lower than the plate voltage divided by 20, in order to assure proximity to cut-off for anode bend plate detection. I say "proximity" because some bias current must flow in this circuit with the 5k bias resistor at the cathode if the plate voltage is to exceed 20V.
Enjoy the attachment. Comments invited.
- An_advanced_regenerative_circuit_s1.gif (111 KB)
- An_advanced_regenerative_circuit_s2.gif (122 KB)
- An_advanced_regenerative_circuit_s3.gif (39 KB)
Dear RM friends, Prof Rudolph has correctly pointed out the following in a private email:
When you describe the "advanced regenerative circuit 1953", you mention:
"The author replaced the grid detection with by anode anode bent
From the schematic in the article of Irving Gottlieb the 1st triode
(6SN7)/2 does not have an AF path at its anode to subsequent stages. So
I cannot see how anode bent detection can happen.
I only see a cathode follower with the first triode and a grid detector
(with feedback to the resonant circuit) with the 2nd one. Grid current
will flow within the 2nd triode, but this will not load the resonant
Can you please clarify these discrepances? Thank you.
Prof. Rudolph's interpretation is indeed correct.
Perhaps I can explain where my interpretation went wrong, as this may serve as a good lesson.
The biggest problem was simply a visual mistake following the audio signal. It seemed that the audio signal was being coupled from the plate of the first triode to the grid of the audio preamp. This is clearly not correct, as the plate signal at the first triode is well bypassed at audio and RF frequencies, and can't reach the grid of the audio preamp.
The second triode plate is starved with a 470kOhm load and has nearly zero grid bias with 1.5Meg to ground with R4. The plate voltage should be around +15V, as can be determined from the following load line graph.
This low plate voltage is ideal for grid leak detection. At this low plate voltage, the grid leak bias should be very close to zero volts, and the grid diode conductance will remain adequate for effective detection. A much higher plate voltage would not let the grid conduct until the grid became substantially positive. A need for low plate voltage to maintain grid detection efficiency is also the reason why battery sets from the 1920's usually had a 22.5V supply tap for the grid leak detector, while other tubes ran with higher plate voltages.
After my initial fateful mistake of following the audio path incorrectly, I convinced myself that anode bend detection was possible in the first triode. This assessment is also flawed. Ideally, a fixed DC cathode bias would be needed at the cathode, so that the first triode could be run near cut-off with a substancial plate voltage for high plate detection efficiency and output voltage range. The 4.7k resistor does not provide a fixed bias because it is not bypassed and because it's drop is 0V when the plate is cut off. A pull-up from the cathode to a positive voltage with a bypass capacitor would be needed to insure a fixed cathode voltage.
I learned several lessons from this correction. Some of the lessons were technical, and another lesson was of very different nature: I should not make posts to RM at 3AM, as a way to cope with insomnia!
Thank you very much Prof Rudolph, for helping to correct the understanding of this circuit.
Forum Readers, while surveying various post-war radio designs with the German military surplus RV 12 P2000 tube, I came across this interesting regenerative detector design, that includes negative audio feedback from the output speaker to the detection grid. The radio is the Siemens SK468W from 1947-49, and all three tubes in the signal path are the RV 12 P 2000 tube. The first of these is an RF preamp, the second, a regenerative detector and the last the audio output. It uses a RGN364 as the power rectifier.
The follwoing shows schematic detail with the negative audio feedback comming from the speaker to the detection grid with C20=400pF, W13=200k and W5=2Meg W5a=50k.
The schematic does not show any possible adjustment for audio gain or amount of negative audio feedback.
The arrows through the three regenerative feedback band coils L8, L9 and L10 suggest that the ammount of regeneration could be adjusted.
There one RF stage preceeding the schematic above, and it includes the volume control at the antenna circuit.
The negative audio feedback superimposed over the postivive (or regenerative) RF feedback reminds me of the OPAMP (Operational Amplifier) concept. Where external negative feedback reduces a crude, but high, forward gain to a precised closed loop value.
This combination of negative audio feedback with positive RF feedback makes me expect that this radio sounded very nice as compared to other regenerative sets.
If the front end RF preamp, were instead, a pentode Converter, a superhet design could have been realized, where even the regeneration control would be set at the factory and unavailable to the casual user. A super het design with regenerative detection was used in the 1932 Philco 80
The following www.Nostalgiaair.org link includes details about the regeneration adjustment by a qualified technician. Look for page 3-26 under the heading "Philco 80 whistles"
As I continue to think about the detection process at the grid of the regenerative detector Rö2, I thing a basic relationship can be established about the effect of Audio negative feedback on the RF detection.
Let's consider what happens when a modulation peak arrives in the AM signal:
- An RF modulation peak at the control grid of Rö2 will cause the average grid bias voltage to become more negative. This trend is followed at an audio rate.
- The more negative control grid bias at Rö2 causes the plate voltage of Rö2 to rise.
- The control grid of the audio output entode Rö3 follwos this voltage rise, and increases plate current for Rö3.
- The increased plate current at Rö3 causes a drop in voltage (more negative) at the transformer primary,
- If the transformer winding convention is correct, a (positive) rise of voltage will appear at the secondary driving C20.
- Now the Audio negatvie feedback closes via C20 W13 W5 W5a, as this positive voltage rise couples to the control grid of Rö2, which in step 1, was going negative.
This negative feedback trip around the loop is beneficial to reduce distorition in the audio amplification of Rö2, Rö3 the output transformer and speaker.
But there is another effect happening at the control grid of Rö2: The grid, which was originally forward biased by the modulation peak at step 1. is now even more forward biased, by the positive going audio signal of step 6. Remember that the foward bias of the grid at the positive RF peaks is what pushed the average audio grid bias negative.
The converse is also true: a trought in the modulation envelope will tend to cut off grid bias, which will be further reduced by the negative audio feedback.
So we have that the forward conduction of the grid is reinforced, while the reverse grid turn-off is also enhanced. This is helpful to detect weak signals more effectivelly, as grid diode behaviour is reinforced.
In general, regenerative circuits owe much of their sensitivity to small signals from the reinforcement of diode behaviour with RF positive feedback. After followoing the feedback path of the audio signal it is clear that negative audio feedback also reinforces diode behaviour at small signals.
When small RF levels are present at the tube grid, the grid conduction appears quite linear ,like a resistor, so not audio detection occurs; only RF amplifiction is possible, until RF regeneration increases the level at the grid, and negative audio feedback further sharpens diode behaviour at the grid.
I found a similar enhencement of the detection process and improved audio distortion with negative audio feedback in an post I made a few months ago about negative audio feedback to the cathode of the 12AV6 detector-diode/triode-preamp tube. In that thread, I had found the same type of audio negative feedback applied to a diode-triode tube in the The Philips796A from 1936.